← All hardware
214 TOPS of edge vision compute in an M.2 card for under $260.
Pros
- Exceptional TOPS-per-dollar and per-watt for vision
- Compact M.2 form factor fits embedded systems
- Scales to high-density PCIe cards for multi-stream
Cons
- Younger ecosystem and smaller community than NVIDIA/Hailo
- Vision-centric; not aimed at generative AI
- Some SKUs have shown limited availability / pre-order status
✓ Where it shines / best for
- Embedded edge vision in compact/fanless devices
- Upgrading NVRs and industrial PCs with an M.2 AI slot
- Developers prototyping high-efficiency edge inference
✕ Not the best fit for
- Generative AI/LLM training
- Workloads needing multi-card PCIe-scale throughput
- Non-technical users wanting turnkey appliances
Features
- ✓ On-device / offline
- ✓ Edge AI
- ✓ Real-time
- ✓ Computer Vision
- ✓ Low-power / efficient
- ✓ Inference Accelerator
- ✓ M2 Module
- ✓ In Memory Compute
- ✓ SDK
Pricing
| Plan | Price | Billing | Notes |
|---|---|---|---|
| Metis M.2 AI Accelerator Card | ~$149 | one-time | Single Metis AIPU on M.2 2280 module; exact price via Axelera/distributors |
| Metis M.2 Evaluation Kit | Contact vendor | one-time | Dev kit / carrier board bundles priced separately |
Pricing verified from the official source. Prices change often — confirm on the vendor's site before buying.
Specifications
| power | 3.5-9W |
| memory | 1GB DRAM (Max up to 16GB) |
| interface | PCIe Gen3 x4 (M.2) |
| architecture | Quad-core Metis AIPU (digital in-memory compute) |
| ai_performance | 214 INT8 TOPS (M.2); up to 856 TOPS (quad PCIe) |
Sponsored
A full review is being generated for this product and will appear here shortly.